Question

我要试着说清楚。

首先是表格：

CARROSSERIE

MARQ      MODEL     SILHOUETTE   ID

citroen   c3        coupe        1
citroen   c3        sport        2
citroen   c4        coupe        3
citroen   c4        sport        4
acura     cdx       cuv          5
...       ...       ...          ...

table2与table1具有相同的字段，但是有更多字段（只有1个字段很有趣）

alltable

 SAME AS TABLE1      zone      
 ...                 EUR
 ...                 EUR
 ...                 USA
 ...                 RUS
 ...                 CHI
 ...                 ...

所以我只是简单地将table1和table2加入到他们的相似字段中以获得＆＃34; zone＆＃34;来自table2。

以下是查询：

 SELECT C.model_marq AS model_marq, C.model_name AS model_name, C.silhouette AS silhouette, IIF(T.ZONE IS NULL, 'ROW', T.ZONE) AS zone
    INTO zone_vehicule
    FROM CARROSSERIE C
    LEFT JOIN alltable T
    ON C.model_marq= T.MARQUE
    AND C.model_name = T.MODELE
    AND C.silhouette = T.CARROS
    GROUP BY model_marq, model_name, silhouette, zone

我获得了良好的结果，因为我得到了所有的＆＃34;区域＆＃34;来自table2，取决于table1的字段。问题是我需要在这个查询中处理一些事情：

如果来自table1的汽车（字段model_marq，model_name，silhouette）与table2位于7个不同的区域（7个区域实际上是所有可能的区域），我会进入结果集：

MARQ      MODEL     SILHOUETTE   zone

citroen   c3        coupe        EUR
citroen   c3        coupe        CHIN
citroen   c3        coupe        ASI
citroen   c3        coupe        RUS
citroen   c3        coupe        AML
...       ...       ...          etc(all the 7 zones)

这是完全正常的查询我正在使用。我想要的是，如果查询返回这种情况（汽车在所有7个区域中），而不是在结果集中有7行与所有不同的区域，我只想要1行＆＃39; *＆＃39 ;在区域字段中（事实上，只有当它在所有区域中时，它才会合并1行中的7行）

我需要的结果示例：

MARQ      MODEL     SILHOUETTE   zone

citroen   c3        coupe        * (because it was in all the 7 zones)
citroen   c3        sport        CHIN
citroen   c3        sport        RUS
citroen   c3        sport        EUR 
citroen   c4        cuv          AML
acura     cdx       sport        *
acura     cdx       coupe        EUR
...       ...       ...          ...

我不知道从哪里开始。我希望能够明白这一点。随意提问。

Answer 1

我的方法使用子查询来查找列出了7个区域的汽车，然后使用该子查询的结果通知主查询的IIF语句确定“区域”字段。根据您的数据，它会做出一些可能无法证明是正确的假设，例如alltable有carId类似于您需要查看的每辆汽车的ID，以及如果计算每个carId的行数，它将返回不同区域的数量。您可以根据实际情况根据需要调整查询，但希望概念很明确：

SELECT C.model_marq AS model_marq, C.model_name AS model_name, C.silhouette AS silhouette, IIF(T.ZONE IS NULL, 'ROW', iff(allZones.carId is not null, "*",  T.ZONE) AS zone
    INTO zone_vehicule
    FROM (CARROSSERIE C
    LEFT JOIN alltable T
    ON C.model_marq= T.MARQUE
    AND C.model_name = T.MODELE
    AND C.silhouette = T.CARROS) Left join 
    (select distinct carId, count(carId) 
      from alltable 
      group by carId
      Having count(carId) = 7) as allZones on t.carId = allZones.carId
    GROUP BY model_marq, model_name, silhouette, zone

这是子查询：

(select distinct carId, count(carId) 
 from alltable 
 group by carId
 Having count(carId) = 7) as allZones

您可以使用并单独运行以评估结果的准确性。

Answer 2

我在这台PC上没有SQL，但你需要在第一次查询时使用RANK和PARTITION。然后，您可以操纵以获取*记录。

使用RANK / PARTITION的示例：http://www.besttechtools.com/articles/article/sql-rank-functions

Answer 3

从zone删除group by，并在选择列表中使用此表达式：

case count(T.ZONE) when 0 then 'ROW' when 1 then min(T.ZONE) when 7 then '*' end

如果这对Access不起作用，您可以将其转换为：

iif(count(T.ZONE) = 0, 'ROW',
    iif(count(T.ZONE) = 1, min(T.ZONE),
        iif(count(T.ZONE) = 7, '*', '')))

Answer 4

这对我有用，但我觉得它有点复杂，而且我的需要很长。我会尝试使用@TPhe答案来改进它。

SELECT model_marq, model_name, silhouette, zone into ZONE_MODELE FROM
(SELECT C.marque AS model_marq, C.modele AS model_name, C.carros AS silhouette, IIF(T.ZONE IS NULL, IIF(model_name = 'DIVERS', '*', 'ROW'), IIF(C.countzones = 7, '*', T.ZONE)) AS zone, C.countzones as countzones FROM (SELECT marque, modele, carros, count(*) as countzones FROM (SELECT marque, modele, carros, zone from (SELECT CA.model_marq AS marque, CA.model_name AS modele, CA.silhouette AS carros, IIF(TMP.ZONE IS NULL, 'ROW', TMP.ZONE) AS zone FROM CARROSSERIE CA LEFT JOIN alltable TMP ON CA.model_marq = TMP.MARQUE AND CA.model_name = TMP.MODELE AND CA.silhouette = TMP.CARROS GROUP BY CA.model_marq, CA.model_name, CA.silhouette, TMP.ZONE ) group by  marque, modele, carros, zone) a GROUP BY marque, modele, carros) C LEFT JOIN alltable T ON C.marque= T.MARQUE AND C.modele = T.MODELE AND C.carros = T.CARROS GROUP BY C.marque, C.modele, C.carros, zone, countzones) TMP2 GROUP BY model_marq, model_name, silhouette, zone

如果有人知道如何制作相同的简化查询，我将不胜感激

COUNT根据列返回JOIN中的行

4 个答案: